Cambridge Archaeology Field Group July 2020

Cambridge Archaeology Field Group July 2020

Reach Roman Villa, Cambridgeshire 1999 (RVIL99):

The Results of Fieldwalking.

Cambridgeshire HER Event N⁰: ECB6255

Colin Coates for Cambridge Archaeology Field Group, with contributions by Alice

Lyons and Bill Hughes.

The Results of Fieldwalking at Reach, Cambridgeshire (RVIL99).

i

Summary

Cambridge Archaeology Field Group (CAFG) carried out a fieldwalking exercise at the Reach Roman

villa and Iron Age site in 1999, at the request of the Archaeology Field Unit of Cambridgeshire County Council.

This was in order to aid the assessment of the deterioration of the Scheduled Monument due to the impact of

agricultural activities. The fieldwalking area and scheduled site is centred on approximately TL5727565300. It

is enclosed in a triangular area, with the ‘Devil’s Dyke’, an ancient ditch and bank to the northeast, the

Swaffham Prior road to the west and the remains of a disused railway to the south.

The Villa was partly excavated in 1892-3 by T. McKenney Hughes and T. D. Atkinson of the Cambridge

Antiquarian Society. It was described as a corridor villa which was aligned northeast/southwest, with wings at

each end, each having apsidal projections. The southeast wing had a further apse to the southeast and the

remains of a hypocaust. In situ flue tiles suggested that it may have been a bathhouse. These features are

probably indicative of later development of the property, which may have originated as a simpler structure,

perhaps like the aisled phase of the nearby building at Exning, excavated by Ernest Greenfield in 1958 – 59.

For the fieldwalking exercise, the site was marked out with canes in a grid of 10 metre cells. The first strip

started at the junction of the path and Devils Dyke, with the strips aligned at 90 degrees to the Dyke. Strips were

numbered 0 to 41 and cells A to Z and then AA to AG; although the length of some strips was constrained by the

field boundaries.

The finds were washed and sorted with the brick and tile being separated out from the pottery. The

ceramic building material (CBM) finds were assessed by CAFG members, whilst the pottery was submitted to

Alice Lyons (Lyons Archaeology) for identification, dating and cataloguing.

Although little dating evidence was found amongst the CBM, an early phase of building is hinted at by a

small quantity of Iron Age/Early Roman daub, tegula fragments with relatively thicker flanges, a tegula

fragment with an extremely thick bed and perhaps two thicker fragments of imbrex. The Roman CBM was found

in a limited number of fabric types, with almost threequarters by weight, occurring in a single, uniformly well-

fired type.

A small number of possible medieval brick fragments was observed amongst the Post Roman CBM, which

was mostly plain roof tile. The location of this material across the site has the appearance of a typical manuring

distribution.

Plotting the pottery finds’ distribution by era was more enlightening. Early Iron Age pottery was located to

the north-east of the later villa site in a slightly elevated position, near the area of the Devil’s Dyke. Finds

quantities increase through the Iron Age and into the Early Roman era, spreading south-westwards. Pottery

finds then begin to occur around the location of the villa, increasing in intensity through the Romano-British

era. By the Medieval and Post Medieval eras, the site appears to have returned to agricultural use. The finds

density decreases and become much more diffuse across the site, indicative of the typical manuring of arable

farmland.

Although not a grand property, together with the hypocausts and the evidence for distant trading links with

the recovery of Samian ware and Spanish amphorae sherds, the inhabitants appear to have prospered

somewhat, on the edge of the rich fenland zone.


ii

Contents

Section Page

Summary . . . . . . . . . . i

Contents . . . . . . . . . . ii

List of figures . . . . . . . . . . iv

List of tables . . . . . . . . . . v

Introduction . . . . . . . . . . 1

Archaeological background . . . . . . . . 1

Topography and geology . . . . . . . . 1

Fieldwalking methods . . . . . . . . . 2

CBM assessment – Roman . . . . . . . . 2

Introduction . . . . . . . . . . 2

Fabrics . . . . . . . . . . . 3

Roof tile . . . . . . . . . . 5

Tegulae . . . . . . . . . . 5

Imbrices . . . . . . . . . . 7

Tesserae . . . . . . . . . . 8

Box and Combed Tile . . . . . . . . . 9

Plain Tile . . . . . . . . . . 10

Brick . . . . . . . . . . 10

Mortar . . . . . . . . . . 11

Plaster . . . . . . . . . . 11

Daub . . . . . . . . . . 12

Indeterminate . . . . . . . . . . 12

Markings . . . . . . . . . . 12

Discussion . . . . . . . . . . 13

CBM assessment – Post Roman . . . . . . . . 14

Introduction . . . . . . . . . . 14

Brick . . . . . . . . . . 14

Tile . . . . . . . . . . 15

Discussion . . . . . . . . . . 16

Conclusions . . . . . . . . . . 17

Pottery assessment . . . . . . . . . 19

Introduction . . . . . . . . . . 19

Results . . . . . . . . . . 19

Summaries of fabrics and forms . . . . . . . . 19

Prehistoric . . . . . . . . . . 20

Iron Age . . . . . . . . . . 20

Late Iron Age . . . . . . . . . . 20

Late Iron Age/Early Roman . . . . . . . . 20


iii

Section Page

Early Roman . . . . . . . . . . 20

Romano-British . . . . . . . . . . 21

Medieval . . . . . . . . . . 21

Post Medieval . . . . . . . . . . 22

Conclusions . . . . . . . . . . 22

Archiving . . . . . . . . . . 23

Acknowledgements . . . . . . . . . 23

Bibliography . . . . . . . . . . 24

Appendix 1. Extent of fieldwalking grid showing Roman CBM finds spots . . . 26

Appendix 2. Extent of fieldwalking grid showing Post Roman CBM finds spots . . 27

Appendix 3. Pottery forms by era . . . . . . . 28

Appendix 4. Pottery finds distributions by era, per 10m gridsquare . . . 30

Iron Age . . . . . . . . . . 30

Late Iron Age . . . . . . . . . . 30

Late Iron Age/Early Roman . . . . . . . . 31

Early Roman . . . . . . . . . . 31

Romano-British . . . . . . . . . 32

Medieval . . . . . . . . . . 32

Post Medieval . . . . . . . . . 33

Appendix 5. Gridsquares by finds bags table . . . . . . 34


iv

List of figures

Figure Page

1. Reach villa excavation plan . . . . . . . . 1

2. Fieldwalking area map . . . . . . . . 2

3. Reach Villa Tegula flange profiles . . . . . . . 5

4. CBM category Tegula, dimensions . . . . . . . 6

5. CBM category Imbrex, thickness by frequency . . . . . 7

6. Imbrex fragment (Bag 91) with 'XX' marks . . . . . . 7

7. Tesserae Aspect Ratio by frequency . . . . . . . 8

8. Tesserae Aspect Ratio by surface colour . . . . . . 9

9. Combed tile Bag 110 . . . . . . . . . 9

10. Combed tile Bag 117 . . . . . . . . 9

11. CBM category Tile, thickness by frequency . . . . . . 10

12. Post Roman peg tile (Bag 258) . . . . . . . 16

13. Post Roman Plain roof tile, frequency by thickness . . . . . 16

14. Reach landscape elevation profile . . . . . . . 17


v

List of tables

Table Page

1. Summary of Roman CBM forms by fabric type . . . . . . 3

2. Summary of Roman CBM fabric types . . . . . . . 4

3. Summary of Roman mortar fabrics . . . . . . . 11

4. Summary of Roman plaster fabrics . . . . . . . 11

5. Summary of daub fabrics . . . . . . . . 12

6. Summary of Roman CBM markings . . . . . . . 12

7. Roman CBM with mortar on broken edges . . . . . . 13

8. Medieval CBM fabric summary . . . . . . . 14

9. Post Medieval CBM fabric summary (after Lyons) . . . . . 14

10. Post Roman brick fabrics . . . . . . . . 15

11. Post Roman tile fabrics . . . . . . . . 15

12. Pot sherd quantities per era (adapted from Lyons 2019) . . . . 19

13. Sherds without gridsquare locations . . . . . . . 19

14. Late Iron Age pottery fabric summary . . . . . . 20

15. Late Iron Age/Early Roman pottery fabric summary . . . . . 20

16. Early Roman pottery fabric summary . . . . . . 21

17. Romano-British pottery fabric summary . . . . . . 21

18. Medieval pottery fabric summary . . . . . . . 22

19. Post Medieval pottery fabric summary . . . . . . 22


1

Introduction

Cambridge Archaeology Field Group

(CAFG) carried out a fieldwalking exercise at

the Reach Villa site in 1999, at the request of

the Archaeology Field Unit of Cambridgeshire

County Council. This was in order to aid the

assessment of the deterioration of the

Scheduled Monument (NHLE: 1006875), due

to the impact of agricultural activities.

An initial analysis by Bill Hughes (CAFG

1999) of the finds by category; Roman Tile,

Roman Brick, Post Roman Tile, and Post

Roman Brick classified by Number of

finds, Total weight, Maximum weight,

Average weight, Minimum weight, per find

location showed that dispersal of material had

occurred, but when this took place could not

be determined. This report will concentrate on

the more detailed assessment of the ceramic

building materials (CBM), utilising the data,

which Bill recorded in a number of

spreadsheets. The pottery was identified,

catalogued and dated by Alice Lyons (Lyons

Archaeology).

Figure 1. Reach villa excavation plan (after Atkinson, 1894).

Archaeological Background

The Villa was partly excavated in 1892-3

by T. McKenney Hughes and T. D. Atkinson

of the Cambridge Antiquarian Society (1894,

229). It was described as a corridor villa,

aligned northeast/southwest, with wings at

each end having apsidal projections (Fig

1).The southeast wing had a further apse to the

southeast and the remains of a hypocaust and

in situ flue tiles suggest that it may have been

a bathhouse (RCHM 1972, 85-90). The wall

lines of the villa are nicely picked out on aerial

photographs of 1977 (CUCAP 1977, CCJ5).

Topography and Geology

The site lies on chalk at approximately 12

metres O.D. on ground sloping gently South-

West to the Fen edge. It is enclosed in a

triangular area, with the ‘Devil’s Dyke’, an


2

Figure 2. Fieldwalking area map.

ancient ditch and bank to the northeast, the

Swaffham Prior road to the west and a disused

railway to the south (Fig 2).

Fieldwalking methods

For the fieldwalking exercise, the site was

marked out with canes in 10 metre wide strips

subdivided into 10 metre by 10 metre cells.

The first strip started at the junction of the path

and Devils Dyke at National Grid reference

TL 57502 65275, with the strips aligned at 90

degrees to the Dyke. Strips were numbered 0

to 41 and cells A to Z and then AA to AG:

although the length of some strips was

constrained by the field boundaries.

Each cell was walked by one person; all

finds from each cell being picked up and

placed along with a tag, having a unique finds

number, in a plastic bag. Some cells however,

produced more than one bag of finds. Such

bags were designated with the same finds

number, but were additionally marked 1 of 2, 2

of 2 etc; five squares received two finds

numbers however. Following washing, the

finds were sorted and the brick and tile

separated out from the pottery. The CBM finds

were assessed by CAFG members, whilst the

pottery was submitted to Alice Lyons (Lyons

Archaeology) for dating, identification and

cataloguing.

CBM assessment – Roman

Introduction

Fragments of Roman ceramic building

materials (CBM) were recorded by weight and

firing grade, inclusions were noted and colour

of fabrics was determined in accordance with

the Munsell Soil Color system. The data were

recorded in a Microsoft Excel spreadsheet and

analysed with the aid of a Pivot Table (Table

1). Measurements were made to the nearest

millimetre unless indicated otherwise. The

daub, mortar and plaster, are dealt with later.


3

FABRICS

F1/F1a F2 F3 F4 F5 F6 F7 TOTALS

CBM TYPE N° WT(g) N° WT(g) N° WT(g) N° WT(g) N° WT(g) N° WT(g) N° WT(g) N° WT(g)

BOX 6 569

6 569

BRICK 11 3766 2 824 2 685

3 372

18 5647

COMBED 8 315 1 91 2 186

11 592

IMBREX 36 1769 5 191

1 35

42 1995

INDET - - - - - - - - - - - - - - 644 9677

TEGULA 21 1594 1 166 3 888

1 462

26 3110

TESSERA 64 1601 1 30 6 138

71 1769

TILE 121 9285 9 408 24 1648 5 502 4 475 2 167 165 12467

TOTALS 267 18899 19 1710 37 3545 6 537 4 834 4 475 2 167 983 35826

Table 1. Summary of Roman CBM forms by fabric type.

The assemblage was examined by 10x

magnification hand lens in order to aid the

compilation of a catalogue of fabric types

(Table 2). The forms of bricks and tiles were

determined where possible and by reference to

Brodribb (1987). Representative samples of

forms and fabrics were retained, with the

remainder and unidentifiable fragments being

disposed of after recording. The percentage

weights in Table 2 have been calculated for the

identifiable fragments only.

Unusual features, where observed, such as

marks made by humans, fabric colours and

inclusions were recorded. The cross-sections

of tegula flanges were drawn, including in

particular any identifiable cutaways which

were categorised according to Warry (2006a).

Any evidence for how the flanges may have

been formed was also recorded. A plan of the

fieldwalked grid with Roman CBM find spots

can be found in appendix 1.

The site codes referred to in this report,

including that for the Reach Villa

investigation, are internal CAFG codes.

Although most of the identifiable post

Roman brick and tile was thought to have been

removed from the assemblage during the

original sorting process, a quantity was

subsequently identified and is dealt with

separately.

The Munsell Soil Color ranges referred to

in this assessment are: 2.5YR

(pinkish/orange), 5YR (light red/orange),

7.5YR (dark red/orange), 10R (red).

Fabrics

The most outstanding feature of the Roman

CBM fabrics from Reach is that they are

predominantly uniformly well fired, in fine

often micaceous sandy clays. By far the most

common fabric is F1. It contains various

quantities and combinations of red ferruginous

(haematite), calcite and/or quartzite inclusions.

Fine black particles were noted in twelve

examples of F1 fabric, typically in the break.

These are most likely ferruginous (magnetite)

inclusions in the sand. Approximately one

third of the examples had no macroscopically

visible inclusions and these have been

assigned to a subset of the main fabric type

(F1a). All CBM forms are represented in the

F1/1a fabric, with box tiles only appearing in

it.

There is a further smaller subset of

fragments in a similar fabric (F2), which have


4

additional silty clay pellets, patches and

streaks; all forms appear in this fabric. This

fabric type has been observed at Harlton

(Coates 2015), where it was the dominant

fabric.

The fourth major fabric group (F3), is

generally not as well fired as the preceding

ones. Fragments of it typically have heavily

reduced cores, while four examples have fine

micaceous inclusions and one, fine black

inclusions, as in fabric F1.

The remaining three groups represent much

poorer fabrics. Those of F4, have streaky,

red/brown cores, which may be indicative of

poor mixing of the tempering sand with the

clay and/or low firing temperatures. With the

exception of one possible fragment of imbrex,

they are all plain tiles of various thicknesses.

The remaining three small fabric groups

contain examples of CBM fragments which

have characteristics that do not readily fit into

any of the foregoing groups. Three brick and

one tegula fragment comprise the members of

fabric F5. Their fabric cores are very ‘lumpy’

in appearance in the break, having large

granular size. The four tile fragments of F6 all

displayed voids in their cores. The three

thicker examples (17-25mm), could have been

fragments of tegulae which occasionally

contain voids, where overlapping layers of

clay have not been well amalgamated. Finally,

the two fragments of tile in F7 were overfired

and/or burnt, but it was not possible to be sure

if this was post manufacture or not.

FABRIC N° WT(g) %Wt DESCRIPTION

F1 198 12814 49 UNIFORMLY WELL FIRED, SANDY CLAY WITH; FERRITIC (150 EXAMPLES), CALCITIC (56) AND/OR QUARTZITE (112) INCLUSIONS. SURFACE COLOUR RANGES: 10R (43), 2.5YR (117), 5YR (21), 7.5YR (10), UNRECORDED (7)

F1a 69 6067 23.2 UNIFORMLY WELL FIRED, SANDY CLAY WITH NO INCLUSIONS VISIBLE MACROSCOPICALLY (70 EXAMPLES). SURFACE COLOUR RANGES:10R (19), 2.5YR (38), 5YR (3), 7.5YR (2), UNRECORDED (7)

F 2 22 1710 6.5 GENERALLY UNIFORMLY WELL FIRED, SANDY CLAY WITH SILTY CLAY PELLETS, PATCHES AND/OR STREAKS (24 EXAMPLES). SURFACE COLOUR RANGES:10R (5), 2.5YR (6), 5YR (4), 7.5YR (4), UNRECORDED (3)

F3 37 3545 13.6

GENERALLY NOT UNIFORMLY WELL FIRED, WITH REDUCED BANDS IN THE CORE. SANDY CLAY WITH; FERRITIC (20 EXAMPLES), CALCITIC (8) AND/OR QUARTZITE (10) INCLUSIONS. SURFACE COLOUR RANGES: 10R (12), 2.5YR (12), 5YR (10), 7.5YR (2) , UNRECORDED (1)

F4 7 537 2.1

NOT UNIFORMLY WELL FIRED, WITH REDUCED, STREAKY RED/BROWN BANDS IN THE CORE. SANDY CLAY WITH; FERRITIC (2 EXAMPLES) AND CALCITIC (8) INCLUSIONS. SURFACE COLOUR RANGES: 10R (3), 2.5YR (1), 5YR (1), 7.5YR (1), UNRECORDED (1)

F5 4 834 3.2 A POOR, ‘LUMPY’ FABRIC, SANDY CLAY WITH; FERRITIC (2 EXAMPLES) AND CALCITIC (3) INCLUSIONS. SURFACE COLOUR RANGES: 10R (3), 2.5YR (1)

F6 4 475 1.8 GENERALLY NOT UNIFORMLY WELL FIRED, VOIDS IN THE CORE. SANDY CLAY WITH; FERRITIC (1 EXAMPLE) INCLUSIONS. SURFACE COLOUR RANGE: 2.5YR (4)

F7 2 167 0.6 POOR, OVERFIRED/BURNT. FERRITIC, CALCITIC AND QUARTZITE INCLUSIONS (1 EXAMPLE). SURFACE COLOUR RANGE: 10R (1), UNRECORDED (1)

Table 2. Summary of Roman CBM fabric types

As can be seen from Table 2 above, the

Reach Villa CBM are dominated by one

fabric; F1. When taken together with the

similar F1a fabric, the two probably represent

different parts of a single continuum of

variability in clay source and added sand

temper. They make up almost three quarters by

weight of the entire Roman CBM assemblage.


5

This lack of variability in fabric types is not

unusual for Roman sites in Cambridgeshire,

which has been noted at for instance Harlton

(Coates 2015) and Melbourne (AOC 2017).

Roof Tile - Tegulae

01 02 03 04 05 06

07 08 09 10 11 12 Figure 3. Reach Villa tegula flange profiles. (Scale 1:2) Broken/damaged

All the flange profiles illustrated in Fig.3

above have been drawn as if left handed to aid

comparison.

There appears to be little variety in the

tegula flange types amongst the Reach Villa

assemblage: a number of them sharing some

common characteristics. Several for instance,

have been trimmed at the outer base of the

flange (Fig.3:01, 03, 04, 07, 08, 09, 10), a trait

seen on tiles from other Cambridgeshire sites,

for example Great Eversden (CAFG

forthcoming a). None of them however,

exhibit the double finger smoothing channels

seen on the inner flange faces of Great

Eversden or Haslingfield (CAFG forthcoming

b) tiles.

A variety of moulds appear to have been

used, including those with vertical sides

(Fig.3:03, 08, 09, 12), inclined inwards

(Fig.3:02, 06?), inclined outwards (Fig.3:01,

07, 10, 11). A mould with inclined sides may

have been more easily lifted away from a still

wet tile although, presumably, a mould with

sides which are inclined inwards would have

been used to produce an inverted tile, with an

insert on the baseboard, (possibly Fig.3:01,

07). Alternatively, a mould with detachable

sides could have been employed (Warry

2006a).

Another of the Reach tegula may have been

produced using an inverted mould (Fig.3:08).

This displays a characteristically flat top to its

flange.

Several of the tegula flanges (Fig.3:03, 05,

06, 11) exhibit vertical lips on their upper

outer edges, which may be indicative of them

being smoothed against the inner face of the

mould.

Only one of the Reach Villa tegula flanges

examined (Fig.3:03), could have carried an

example of a Warry (2006b) type B6 lower

cutaway. However, the cut is quite shallow

compared to other examples and may be no

more than excessive trimming of the lower tile

edge.

Where they reliably survived, up to three

measurements were taken for each tegula

flange; the bed thickness where it meets the

flange, the overall external flange height and

the flange width, measured along a horizontal


6

line from the outer flange face at the height of

the bed, to a point where it would intersect a

line projected down the inside face of the

flange. Statistical analysis of such a small

sample may not give very meaningful results,

unless there are some very strong underlying

factors. Nevertheless, some of the statistics

derived from the Reach villa tegulae fragments

are worth noting.

Figure 4. CBM category Tegula dimensions.

The average flange width of the 12

measurable fragments is 26.3mm, with 6

(50%), falling within one standard deviation.

The distribution of widths spans the range of

16-33mm. This observation is perhaps

unsurprising, as usually the widths of flanges

increased, tapering from upper to lower ends.

One example (Bag 127) demonstrates this, as

its flange width tapers from 29-33mm over a

length of 198mm.

Flange heights vary between 45-51mm

with a median value of 48mm. This small

variation in flange heights, +/-3mm (6.25%),

might suggest that standardised box moulds

were employed in the production of the

tegulae. The small variation in height then,

may be attributable to the differential

shrinkage of the clay. The sample of

measureable flanges however, is perhaps too

small to draw conclusions from with

confidence.

Bed thickness is strongly concentrated

between 16-24mm. The average thickness

being 20.5mm, with 13 (68.4%) examples

falling within one Standard Deviation. There

may be a second population with a thickness

of 29mm, although the measurements having

been made close to the flanges, may not be

representative of bed thickness as a whole.

Only one example of an upper cutaway was

observed, from bag 117. It came from the left

hand corner of a tegula, the flange being

removed with a knife or tool. Although

WIDTH BED HEIGHT Count n 12 19 7

Average 26.3mm 20.5mm 47.7mm 1 Std Dev 5.7mm 3.4mm 2.4mm n @ SD 6 (50%) 13 (68.4%) 3 (42.9%)


7

incomplete, the cut survived for a length of

40mm horizontally, commencing 8mm above

the bed and sloping down towards the top edge

of the tegula, before joing the bed, 7mm from

the edge.

It can be difficult to draw conclusions from

randomly broken sections of tegula flanges

unless the ends of the tiles are present: few

were apparent amongst the Reach assemblage.

Nevertheless, the flanges of four of the tegula

fragments (Fig 3:02, 03, 04, 09), at 32-33mm,

are amongst the widest of those found in

Cambridgeshire in a survey of Roman roofing

tiles by the author (Coates 2014).

Roof Tile - Imbrices

A total of 42 imbrex fragments were

identified amongst the CBM assemblage. Most

were relatively small and abraded. All had

been produced on a sanded former. The outer

surfaces of most of the imbrex

Figure 5. CBM category Imbrex, thickness by frequency.

fragments had been well smoothed

longitudinally; however one was smoothed

transversally whilst another displayed

transverse finger marks. Surface colours were

predominantly in the 2.5YR range (25

examples), with cores fired to the same colour

range (25), with a smaller group (12) with

surfaces in the 10R range and cores likewise

(11). Other than the tile maker’s finger marks,

only one piece had any distinctive marks,

which were noted on an example from Bag 91

(Fig. 6).

Figure 6. Imbrex fragment with 'XX' marks (Bag 91).

Imbrex thickness ranged between 10-20mm,

with a strong peak at 15mm, whilst 27 of the

Count n=41

Average=14.6mm

1 Std Dev=2.26mm

n @ SD=27 (65.9%)


8

fragments (65.9%) fell within one standard

deviation of the average of 14.6mm (Fig. 5).

One fragment was not measureable and it is

possible that the two examples with thickness

of 20mm were actually pieces of ridge tile, as

observed by Brown (1994, 83) for example at

Harold.

Tesserae

Although no fine stone tesserae were

recovered during the fieldwalking, 71 pieces

of tile weighing 1.77kg (6.7% of the Roman

CBM by weight), were found to have been

used, or reused as tesserae. This was

evidenced by mortar surviving on up to five

faces. The majority of the tesserae had been

made from plain tile. However, one was found

to have been part of a combed tile, whilst a

second tessera had what appeared to be part of

a finger mark, indicating that it may have

originally been part of a tegula.

Analysis of the aspect ratio (shorter

side/longer side) of the tesserae, revealed that

the group was predominately sub-square

having a mean ratio of 0.85. Only 4 of the

assemblage were almost exactly square. The

spread of sizes appeared to be reasonably

random, as might be expected from a hand-

made product, although the plot of tessera

aspect ratio by frequency, hints at there being

two peaks in the size profile (Fig.7).

Figure 7. Tesserae Aspect Ratio by frequency.

Plotting tessera aspect ratio by surface

colour (Fig.8) shows a little more detail.

Amongst the 71 tesserae whose colour

attributes were recorded, 3 distinct surface

colour groups were noted. By far the greatest

number of tesserae exhibited a surface colour

in the 2.5YR range. Although the aspect ratio

distribution of the 2.5YR group spans the full

range of values, there may be two distinct

subgroups; one in the range 0.59-0.81, the

other 0.83-1.0.

The second most populous tessera surface

colour group (10R), with the exception of two

Count n=71

Average=0.85

1 Std Dev=0.1

n @ SD=46 (64.8%)


9

outliers, have values of aspect ratio between

0.74-0.93. Thirdly, tesserae with surface

colours in the 5YR group, with the exception

of one outlier, are more tightly grouped than

the other two groups in the range 0.84-0.97.

Figure 8. Tessera Aspect Ratio by surface colour.

Box and Combed Tile

Only six pieces of CBM were identified

as being parts of box tiles. This was largely

due to the partial survival of their corners.

They were all uniformly well fired, with traces

of fine moulding sand on their inner faces.

Figure 9. Combed tile Bag 110

There were a further eleven fragments of

combed tile in the assemblage. They were

predominantly of the uniformly well fired,

sandy F1 fabric, with surface colour ranges;

2.5YR (5 examples), 5YR (3), 7.5YR (1), 10R

(2). Most of the combings were apparently

executed by the use of a worn wooden comb,

with up to six linear marks being visible (Fig

9). One example (Fig.10) however, had two

Figure 10. Combed Tile Bag 117.


10

sets of fine combings at right angles. One

piece of combed tile had been reused as a

tessera.

Plain Tile

Fragments of CBM were assigned as tile,

based on their having at least two intact

surfaces, reliably measureable thickness, but

no other characteristics which could allow

them to be placed in any other CBM category.

In total 164 examples weighing 12.47kg

(47.68% of the Roman CBM by weight), were

recorded. Three fragments from bag 111 were

able to be refitted. The thickness of two

fragments could not be reliably measured.

Figure 11. CBM category Tile thickness by frequency.

The histogram of tile thickness by

frequency (Fig 11), shows a strong grouping of

tile fragments with thickness between 11mm–

22mm, with a possible second peak at 29mm–

30mm. Although the thinner examples could

be parts of box or flue tiles, 99 examples fall

within the range of tegula bed thickness (16-

29mm). The corners of 8 plain tiles had

survived.

Almost threequarters of the plain tile

(73.2% numerically), was produced in the

F1/F1a fabric. The mean thickness of the F1

fabric tile was 18.26mm, with 86 examples,

falling within 1 Standard Deviation, between

13-23mm. Of the remaining fabrics, 23

examples (20.1%), came from F2 and F3.

Brick

Fragments of CBM with thickness greater

than 30mm were designated as brick. There

was a total of 18 such fragments, weighing

5.65kg (21.6% of the Roman CBM by weight).

The average weight was 313.7g, with

thickness in the range 31mm–47mm.

The majority of brick fragments (16), had

thickness in the range 31-37mm, the remaining

two were 46 and 47mm thick respectively.

Several of the fragments with thickness in the

All Tile Count n=164

Average=18.11mm

1 Std Dev=5.22mm

n @ SD=117 (71%)

F1 Fabric Count n=121

Average=18.26mm

1 Std Dev=5.01mm

n @ SD=86 (71%)


11

range 31mm–33mm however, could be

examples of tegulae of extreme bed thickness.

Examples in the F1/F1a fabric were most

common (11examples), with fabrics F2 (2), F3

(2) and F5 (3), also being represented. Surface

colour ranges of the 17 examples recorded

were; 2.5YR (8), 10R (6), 5YR (2), 7.5YR (1).

Mortar was found adhering to the broken

edges of 4 examples; F1a (3), F3 (1).

Mortar

Besides that adhering to a number of pieces

of CBM, 5 pieces of mortar were collected

during the fieldwalking exercise weighing a

total of 84g. They can be categorised as two

distinct types as in Table 3 below.

FABRIC N° TOTAL WT(g)

DESCRIPTION

M1 3 16 OCCASIONAL (1.5mm) TO ABUNDANT (1.0mm) QUARTZ SAND. RARE FERROUS (1.0mm) AND CHALK (3mm) INCLUSIONS (1 EXAMPLE). 2.5Y COLOUR RANGE

M2 2 68 CREAMY WHITE, COARSE MORTAR, ABUNDANT QUARTZ SAND (<3mm), OCCASIONAL BLACK INCLUSIONS (<2mm), FLINT FLAKE (4*5mm), (1 EXAMPLE). 5Y COLOUR RANGE

Table 3. Summary of Roman mortar fabrics.

Plaster

One fragment of plaster came from the

main CBM assemblage, while a further 17

were included in the pottery assemblage

assessed by Alice Lyons. For comparison

purposes, they have been divided into four

fabric types as shown in table 4.

One face of each of two P1 fragments was

painted dark red/brown (5YR/5/6). Abundant,

very fine quartz or mica grains could be seen

glinting through the slightly abraded painted

surface.

The painted layer of a third P1 example

was very similar to that of the others in the P1

fabric, with patches of dark red/brown

(5YR/5/6) over a light pink (5YR/8/4)

undercoat. The surface layer of fine plaster sat

on a much coarser substrate containing sub-

rounded stones <5mm. Twelve examples of P2

fabric were very coarse containing stones

<5mm. They all probably represent substrate

layers similar to that of the P1 example, with

which they were collected from the same

gridsquare.

Two examples in P2 had a pink-stained

surface, one of which carried the negative

impression of fine combing and they had

presumably been applied to brick or tile

surfaces. One example of P2 displayed

impressions of fragments of wood or organic

material in the break.

The single example of P3 was a very fine,

sandy, light beige fabric. It had a light

red/brown (5Y/7/3) painted surface and

organic impressions were noted in the break.


DESCRIPTION

P1 3 25 CREAMY WHITE (2.5Y/8/3), COMMON TO ABUNDANT DARK FERROUS INCLUSIONS <0.25mm, ABUNDANT VERY FINE < 2mm QUARTZ SAND GRAINS, RARE CALCITE <2mm. BASE LAYER COARSE, STONES <5mm.

P2 13 173 PALE CREAMY WHITE (5Y/8/2), OCCASIONAL <0.25mm TO COMMON <1.0mm FERROUS INCLUSIONS, ABUNDANT QUARTZ AND STONES 1-5mm

P3 1 14 LIGHT BEIGE (2.5Y/8/4), OCCASIONAL FERROUS INCLUSIONS <0.25mm

Table 4. Summary of Roman plaster fabrics.


12

Daub

One fragment of daub was observed in the

main Roman CBM assemblage, while a further

five were identified amongst the pottery; all

being very abraded. Only three examples

could be dated with confidence and were

assigned to the Late Iron Age/Early Roman

(LIA/ER) era. Their fabrics are as summarised

in table 5 below.


DESCRIPTION

D1 1 16 UNIFORM, 5YR/7/4

D2 3 6 COMMON CALCITIC INCLUSIONS <0.5mm, ONE FRAGMENT HAS OCCASIONAL QUARTZ <3mm, 10YR/7/4

Table 5. Summary of daub fabrics.

Indeterminate

A large number of CBM fragments were

too abraded or damaged to adequately

determine their form, although their eras of

origin were noted. They were recorded as

indeterminate (INDET) before being

discarded. This needs to be born in mind when

considering the analysis. There were 644

fragments (65.34% numerically), weighing

9.67kg (27% of the total Roman CBM by

weight), with an average weight per sherd of

15.02g.

Markings

As noted above, one fragment of imbrex

had been marked with a double ‘X’. These

marks may have been a tile maker’s tally

BAG N° CBM TYPE DESCRIPTION

110 TEGULA FINGER IMPRESSION ON THE BOTTOM, LOWER EDGE OF THE FLANGE, INDICATING THAT IT HAD BEEN HANDLED WHILE STILL WET.

117a TILE A LIGHT, APPROXIMATELY 4MM WIDE LINEAR MARK, APPARENTLY MADE WITH A NOTCHED/WORN WOODEN TOOL OR STICK. VERY ABRADED.

117b TILE 2 DIVERGING 5mm WIDE MARKS, APPARENTLY MADE WITH A NOTCHED/WORN WOODEN TOOL OR STICK.

126 TILE TWO DIVERGING, SINGLE CURVING FINGER MARKS, WORN/ABRADED, APPROXIMATELY 7MM WIDE.

128a TILE A LIGHT, 4MM WIDE LINEAR MARK, APPARENTLY MADE WITH A NOTCHED/WORN WOODEN TOOL OR STICK, POSSIBLY THE SAME AS 117.

128b TILE A LIGHT, 4MM WIDE LINEAR MARK, APPARENTLY MADE WITH A NOTCHED/WORN WOODEN TOOL OR STICK, POSSIBLY THE SAME AS 117.

130a TILE A LINEAR, DOUBLE FINGER MARK NEAR BROKEN EDGE OF TILE, APPROXIMATELY 5MM AND 9MM WIDE RESPECTIVELY. CONSISTENT WITH BEING MADE BY A MIDDLE AND ADJACENT FINGER.

130b TILE A DOUBLE CURVING FINGER MARK NEAR BROKEN EDGE OF TILE, APPROXIMATELY 7MM AND 9MM WIDE RESPECTIVELY. CONSISTENT WITH BEING MADE BY A MIDDLE AND ADJACENT FINGER.

150 TILE A LIGHT, 4MM WIDE LINEAR MARK, APPARENTLY MADE WITH A NOTCHED/WORN WOODEN TOOL OR STICK, POSSIBLY THE SAME AS 117. THE MARK IS ALIGNED AT APPROXIMATELY 33 DEGREES TO THE SURVIVING EDGE OF THE TILE.

181 TILE A LINEAR, DOUBLE FINGER MARK, WHICH APPEARED TO BE CONVERGING WITH A SINGLE, LINEAR, NOTCHED/WORN WOODEN TOOL MARK, POSSIBLY THE SAME AS 117.

232 TILE SEVERAL SEMI-CIRCULAR MARKS, 4-4.5MM DIAMETER, WHICH MIGHT HAVE BEEN MADE BY FINGERNAILS OR A TOOL.

388 TILE A LIGHTLY DRAWN, LINEAR FINGER MARK OF 5MM WIDTH, ALTHOUGH IT TAPERS OUT A LITTLE.

Table 6. Summary of Roman CBM markings.

marks or used for some other accounting, or

identification purpose. Similar examples have

been noted from York (McComish 2012 177,

209).


13

One piece of combed tile had been reused

as a tessera, whilst a second tessera had what

appeared to be part of a finger mark, indicating

that it may have originally been part of a

tegula. A list of other CBM fragments which

had distinctive markings, is given in table 6.

Discussion.

Excluding the daub, plaster and mortar, a

total of 983 Roman CBM fragments, weighing

35.83kg were collected during the

fieldwalking exercise, many being highly

abraded. Roman CBM was recovered from

157 of the 10m grid squares, at an average

density of 2.28g/m2 for the productive squares.

The ratio of tegula fragments to imbrices

by weight, amongst those which were

definitely identifiable in the assemblage, is a

little low at 1.56, even discounting those

fragments with mortar on broken edges, this

only rises to 1.62. If the weight of plain tile

fragments in the range of tegula bed thickness

(16mm-29mm) is added to that of the

identified tegula fragments, a ratio of 6.26 is

produced, which is unfeasibly high. However,

following Warry (2010, 1), if both imbrices

and tegulae were subject to the same history of

destruction, then the ratio of the average

weight of the resulting fragments, might then

be close to the ratio of their original weights.

For the Reach CBM we arrive at a ratio of

3.65. This figure is a little on the high side for

the suggested range for a standard Roman roof

(Brodribb 1987, 11-12; Ramos Sáinz 2003),

but not excessively so. This might suggest that

the tegulae and imbrices recovered represent

part of a fallen roof, although the excavators

did not mention any evidence for the same.

As noted above, a small number of CBM

fragments had mortar adhering to broken

edges. This suggests that they were reused

from an earlier phase of building on the site, or

imported for general building purposes (Table

7).

BOX BRICK COMBED IMBREX TEGULA TILE

1 2 1 6 3 13

Table 7. Roman CBM with mortar on broken edges.

There was little dating evidence to be found

amongst the Reach CBM. However, an early

phase of building is hinted at by the tegula

fragments with relatively thicker flanges and

there was also a tegula fragment with an

extremely thick bed and perhaps two thicker

fragments of imbrex. Warry (2006a) for

instance, has proposed that early Roman

roofing tiles began much larger and more

robust, before shrinking in all dimensions.

Additionally, although perhaps of low

statistical significance, the five fragments of

daub, most of which were assigned to the

LIA/ER era, could also be indicative of an

earlier episode of building on the site.

Mckenney Hughes and Atkinson’s

excavation drawing, along with the aerial

photographs, show the Reach building to have

been a symmetrical, winged villa, with

hypocausts under the apsidal room floors.

These features are probably indicative of later

development of the property, which may have

originated as a simpler structure; perhaps like

the aisled phase of the nearby building at

Exning, excavated by Ernest Greenfield in

1958 – 59 (Webster 1988, 41 - 66).

As the CBM collected from the Reach

fieldwalking exercise were not from stratified

deposits and we do not know what happened

to the material from the original excavation, it

is not possible to declare with confidence

whether they represent part of a single, or later

phases of construction.


14

Post Roman CBM

Introduction.

Although it was believed that the Post

Roman CBM was removed at the initial

sorting stage, a small number of fragments

were later identified amongst the Roman

CBM. This may have been due to

misidentification, or uncertainty. Three

fragments could not be clearly dated. A small

number of CBM fragments were identified by

Alice Lyons, amongst the pottery assemblage

submitted for assessment. Of these, six were

categorised as Medieval. All were very

abraded and fragmentary; their fabrics are

summarised in table 8. Only one fragment of

F10, from bag 375, had a recognisable form; a

roof tile, while another from bag 321 in F8,

had a small patch of light grey/brown glaze. A

plan of the fieldwalked grid with Post Roman

CBM find spots can be found in appendix 2.


AVG WT(g)

DESCRIPTION

F8 3 17 5.7 UNIFORMLY WELL FIRED, COMMON TO ABUNDANT DARK FERROUS INCLUSIONS <0.25mm, ABUNDANT VERY FINE QUARTZ SAND GRAINS

F9 1 6 6 UNIFORMLY WELL FIRED, RARE FERROUS INCLUSIONS <0.5mm, VERY FINE QUARTZ SAND GRAINS

F10 2 14 7 UNIFORMLY WELL FIRED, COMMON FERROUS INCLUSIONS <1.0mm, ABUND VERY FINE QUARTZ SAND GRAINS, REDUCED CORE

Table 8. Medieval CBM fabric summary (after Lyons).

Eight CBM fragments were identified as

being Post Medieval. These were also mostly

uniformly well fired. Only one fragment, from

bag 41 in F12, had a recognisable form; the

curved end of a pantile. The single example of

the F13, fabric could be a variant of the F12

fabric. These fabrics are summarised in table

9.


AVG WT(g)

DESCRIPTION

F11 2 12 6 UNIFORMLY WELL FIRED, ABUNDANT VERY FINE QUARTZ SAND GRAINS, COMMON CALCITE INCLUSIONS <0.5mm

F12 5 29 5.8 UNIFORMLY WELL FIRED, RARE TO COMMON FERROUS INCLUSIONS, COMMON TO ABUNDANT VERY FINE QUARTZ SAND GRAINS

F13 1 11 11 WELL FIRED, COMMON FERROUS INCLUSIONS <0.5mm, RARE VERY FINE QUARTZ SAND GRAINS, LIGHT GREY REDUCED CORE

Table 9. Post Medieval CBM fabric summary (after Lyons).

Brick.

Three CBM fragments (Table. 10), were

identified as possibly being the corners of Late

Medieval bricks (Rob Atkins pers. Comm.), in

two fabric types:

F14 - the fabric tends to appear ‘lumpy’ in the

break, having a structure with the appearance

of large granules; one fragment having a large

piece of grog exposed. One fragment (bag

218) had a light creamy slip on one face.

F15 – this is a much more uniformly, well-

fired fabric. The single example of this type

came from bag 155. This fragment represented

the corner of a brick, whose two adjoining top

edges had sunken margins.

Ian Betts has given the simplest and most

reasonable explanation for such features (Betts

1996, 7). The bricks were formed in simple

rectangular moulds. When the mould was

lifted up away from the block of clay, material

was dragged up with it and left standing proud

of the top surface. The brick maker then might

have used the mould to press the excess

material back down. This technique would


15

need to have been applied twice, once each

from diagonally opposite corners, if all four

edges were to be tidied.

Betts (1996, 7) noted that a large number of

red bricks with sunken margins have been

found in London, dating from between the

mid-late 15th century to the 17

th century. In

eastern England, bricks were being traded

through Ely in the 14th century for instance,

notably from kilns at Wisbech, where they

may have arisen due to excess production from

the castle construction works (Sherlock 1999).

The inland port, sitting at the head of Reach

lode, rose to some importance. Although it

was very much in decline by the early 18th

century, goods including bricks were still

being imported there (BHO).


DESCRIPTION

F14 2 382 POORLY FIRED, ‘LUMPY’ FABRIC. OCCASIONAL FERROUS <2mm, 2.5YR/6/6 – 10R/6/4 OR RARE CALCITE <2mm, WELL OXIDISED 5YR/7/6 – 5YR/6/6

F15 1 96 UNIFORMLY WELL FIRED, OCCASIONAL FERROUS INCLUSIONS <5mm, SUNKEN MARGINS ON TWO ADJACENT EDGES. FULLY OXIDISED, 5YR/6/6

Table 10. Post Roman Brick fabrics .

Tile.

With the exception of two possible

fragments of ridge tile, most of the rest of the

Post Roman tile fragments are all plain roof

forms, in two similar sandy fabrics. The cores

of the tiles occasionally have reduced bands

and yellow, red or brown streaks are

occasionally observed in the core. This

colouration may be due to excessive or poorly

mixed sand being added to the clay, or poor

firing temperature control.

With the exception of one example in each

fabric at 22mm thick, their thickness range

spans 9mm - 17mm (Fig 13). Their lower

faces show that they were made on a sanded

bed, whilst their upper faces had a thin, wash

or (self?)slip wiped over them (Carole Fletcher

pers. comm.). The upper faces are pitted

(<1.5mm), which was probably caused by

either the burnout of calcite fragments when

fired, or by leaching out in acidic soil.


DESCRIPTION

F16 10 281 FABRIC APPEARS POROUS. RARE FERROUS <3 - 6mm (2 EXAMPLES), 2.5Y/7/4 –8/4, OCCASIONAL/COMMON CALCITE <3mm (3). 2.5Y/7 – 2.5Y/8(5); 5Y/7 – 8 (5).

F17 59 1015 FABRIC APPEARS POROUS. RARE FERROUS <0.5 -3mm (20 EXAMPLES), RARE TO OCCASIONAL/COMMON CALCITE <0.5 - 4mm (6). 10R/5/4 – 7/4 (16), 2.5YR/6/2 – 7/8 (24), 5YR/6/3 – 7/8 (12), 7.5YR/7/3 – 8/4 (7)

Table 11. Post Roman Tile fabrics.

Examples of F16 fabric type were generally

well fired, with common, fine calcite

inclusions being observed in their cores. Some

are almost entirely fired to a yellow colour

throughout and occasionally pinkish/red

patches may be observed in their cores. The

colour of their wiped surfaces tends to be

yellow or pale greeny/yellow. One example

from bag 258 (Fig 12), had a partial, angled

square peg hole, 12mm across at the top

surface. Slip had flowed into the hole, which

was probably made after the tile was wiped.

The forms of three examples of F16 fabric

could not be determined.


16

Figure 12. Post Roman peg tile (Bag 258).

Common, fine calcite inclusions occur in

their cores, which can appear porous where the

calcite has been leached or burnt out.

CBM fragments in F17 fabric appeared to

be a little more brittle than F16 examples;

showing a tendency for lamina separation.

Their wiped surface colour tends to be a pale

pink. One example from bag 41 has a partial

peg hole where the slip had flowed into the

hole; probably before the hole had been

formed.

Although the post Roman plain roof tile

fragments from Reach can apparently be

divided into two fabric types, in reality, they

may all be members of the same single fabric.

This observation is based on the similarities

between the two fabrics; namely their surface

treatment, the quantities of calcite inclusions

in both and their broadly matching thickness

distributions (Fig 13).

Figure 13. Post Roman Plain Roof Tile, frequency by thickness.

Discussion

The variations in colour may be accounted

for by the differences in temperatures found in

clamp firing for example. As Smith (2001, 35)

has described, products undergoing the full

firing temperature would have attained the

desired yellow colour, whereas those in the

cooler parts, nearer the top of a clamp, would

have fallen further along the colour spectrum,

giving a pinky/red appearance.

F16 F17 Count n = 9 58 Average = 13.1mm 12.6mm 1 Std Dev = 1.73 1.7 n @ SD = 6 (66.7%) 47 (81%)


17

Brick and tile was being imported into

eastern England from Holland from at least the

14th century, with the distinctive ‘small yellow

bricks’ or ‘klinkers’, from at least the 15th

century. The clays used tended to be dredged

from fine, silty riverine deposits, which

produced dense, strong products (Smith 2001,

32-34).

The manufacture of yellow bricks and tiles

commenced indigenously in eastern England

in the mid-18th century and it is most likely

that the examples in the F16 and F17 fabrics

found at Reach date from this period, or a little

later. Bricks and tiles with a pinkish/red or

yellow surface slip or wash and pitted

surfaces, have been found on other

Cambridgeshire sites: the author having seen

examples in similar fabrics from Haslingfield

(CAFG b), Childerley (CAFG c), Ickleton and

Yaxley.

Conclusions

An elevation profile transect, taken across

the Reach landscape from north-east, to south-

west (Fig 14), gives some clues to the reasons

for the pottery concentrations and the villa

being located where they were.

Iron Age and early Roman activity was

largely occuring on and around a slighly

elevated plateau, at approximately 15m O.D.

and may well have extended further to the

east, beyond the Devils Dyke. The later

Roman villa was built on gently sloping land

to the south-west of the higher ground, at

Figure 14. Reach landscape elevation profile.

approximately 12m O.D. Here it had views out

across the fertile fenland landscape, to the west

and north. There has been some conjecture as

to the dating of the Dyke. However, when a

trench was excavated through its earth bank, in

1924 (Fox 1925), Roman era pottery was

discovered on the original land surface below

it. The dyke then, may well have been

constructed in the late Roman, or post-Roman

eras.


18

Although it appears that it was not a vastly

sophisticated property, it did nevertheless have

some distant trading links. There were the

remains of Spanish oil amphorae and a small

amount of Samian pottery.

The nearest major road to Reach during the

Roman era was probably the one some nine

kilometres to the west. It ran from Cambridge

(Duroliponte), northeast towards Thetford and

became known as Akeman Street. However,

reaching this road would have entailed

crossing the river Cam.

A more convenient method for transporting

heavy goods over long distances may have

been the utilisation of one of the other fenland

waterways.

The present village of Reach is connected

to the river Cam by a lode. Lodes are canalised

watercourses, synonymous with the fens.

Although the Fenland waterways have been

much reengineered, particularly during the

Post Medieval era of drainage and

reclamation, the Reach lode was almost

certainly in use during the Roman era.


19

Pottery

Introduction.

At the initial sorting stage of the

fieldwalked assemblage, all of the

recognisable potsherds were removed by

CAFG members and rebagged separately from

the CBM. During the later detailed assessment

of the CBM, a further quantity of potsherds

was identified, which no doubt due to the

highly abraded nature of the assemblage had

been misidentified.

The collection of potsherds itself was

submitted to Alice Lyons (Lyons

Archaeology) for more detailed assessment of

forms, fabrics and dating to major periods

(ERA). From this, a number of sherds were

identified as CBM fragments. These were then

assessed with the main CBM assemblage.

Results.

The results of the pottery assessment were

returned in the form of a Microsoft Excel

spreadsheet, with analysis of the main features

by finds quantities per era, being performed by

the use of a Pivot Table (Table 12).

ERA N° Wt (g) AVG

Wt(g)

PRE 2 9 4.5

IA 13 126 9.7

IA/ESAX 2 31 15.5

LIA 49 431 8.8

LIA/ER 151 1315 8.7

ER 154 800 5.2

RB 164 967 5.9

SAX 1 4 4

LSAX/EMED 1 13 13

MED 63 596 9.5

PMED 26 166 6.4

Totals 626 4458 7.1

Table 12. Quantities by era. (Adapted from Lyons 2019)

Plotting of the finds distributions by era on

the fieldwalking grid was a valuable exercise

(Appendix 4). This showed that activity in the

Iron Age (IA) and Late Iron Age (LIA), took

place to the northeast of the later Roman Villa,

near to the landscape feature known as The

Devil’s Dyke. Two sherds were identified as

Prehistoric (PRE), which were also found in

this area, in grid square K22.

Two sherds fell into a fabric type which is

difficult to identify, especially in view of their

fragmentary and eroded nature; they were

classified as Iron Age/Early Saxon

(IA/ESAX). Only one sherd was definitely

identified as being from the Saxon (SAX) era,

whilst one more, was deemed to span the Late

Saxon/Early Medieval (LSAX/EMED) eras.

Unfortunately, it was not possible to assign

grid square locations for 17 of the pot sherds.

This was due to the incomplete survival of the

original paper records. However, it is not

considered that the results are materially

affected, although some of the percentage era

weights appear high, due to the size of the

sherds (Table 13).

ERA N° Wt(g) % ERA

N° % ERA Wt(g)

%TOTAL Wt(g)

LIA 2 28 4.08 6.50 0.63

LIA/ER 3 14 1.99 1.06 0.31

ER 3 19 1.95 2.38 0.43

RB 5 39 3.05 4.03 0.87

MED 3 25 4.76 4.19 0.56

PMED 1 11 3.85 6.63 0.25

Totals 17 136 3.05

Table 13. Sherds without grid square locations.

Summaries of fabrics and forms.

A copy of the full pottery spreadsheet has

not been included in this report. The quantities

by form for each era were extracted (Appendix

3), with abbreviations explained in the text. A

summary of fabrics and forms follows below.


20

Prehistoric (PRE).

The single fragment of prehistoric pot,

came from a hand made (HM) bowl, in a

reduced fabric tempered with flint

(RW(FLINT)).

Iron Age (IA).

All of the Iron Age pottery, spanning the

C4th BC, to C1st AD was hand made. The

forms were jars and bowls, largely produced in

flint tempered, reduced (RW(FLINT)) and

grey wares (GW(FLINT)). There were also

some jars, in a reduced, shell tempered fabric

(RW(SHELL)) and three fragments of bowl in

a sand tempered ware (STW).

Late Iron Age (LIA).

The Late Iron Age forms (C2nd BC – mid

C2nd AD), were mainly hand made jars or

bowls, although two examples may have been

made on a slow wheel (SW).

FABRIC FAMILY QUANTITY WEIGHT (g)

GW 3 60

GW(CALC) 1 3

GW(FLINT) 2 16

RW(FLINT) 9 94

RW(GROG) 3 27

RW(Q) 3 19

RW(SHELL) 18 110

SGW 8 66

STW 2 36

Table 14. Late Iron Age fabric summary.

Table 14 above, lists the fabric types and

quantities of the Late Iron Age pottery. These

comprise grey wares (GW), reduced wares

(RW), sandy grey wares (SGW) and sand

tempered wares (STW).

Late Iron Age/Early Roman (LIA/ER).

This era covers the period spanning C1st

BC – C2nd AD. Although they are still largely

hand made, more slow wheel and now fast

wheel (WM) examples begin to appear (Table

15). Bowls and jars are dominant, but four

sherds of Spanish olive oil amphora (BAT

AM) are evident.

TYPE FABRIC FAMILY

QUANTITY WEIGHT (g)

HM BAT AM 4 258

HM GW(CALC) 71 597

HM GW(GROG) 6 26

HM OW 1 3

HM OW(SHELL) 1 2

HM RW(Q) 9 72

HM RW(SHELL) 4 19

HM SCW 1 11

HM SGW 20 123

HM SOW 1 48

HM SREDW 2 6

HM STW 8 32

HM/SW GW 1 4

HM/SW RW(Q) 3 34

HM/SW SGW 2 24

SW SGW 9 37

WM GW 1 1

WM SGW 6 13

WM SOW 1 5

Table 15. Late Iron Age/Early Roman fabric summary.

Flint tempering no longer appears amongst

these fabrics; however, there are calcite, grog

and shell tempered fabrics, along with many

more sandy (Q) and sand tempered examples.

Early Roman (ER).

The Early Roman era, spanning the period

C1st AD to C2nd AD, witnesses a much

greater decline in hand made forms, with a

corresponding increase in slow and fast wheel

made products.

Extending the utilitarian range of jar, bowl

and dish forms found in earlier eras, in the

Early Roman era, there are beakers, urns and


21

platters. The jars and bowls themselves

become more flamboyant, with cordons of

burnishing, cross-hatched decoration and

everted rims.

TYPE FABRIC FAMILY

QUANTITY WEIGHT (g)

HM GW(GROG) 2 14

HM RW(SHELL) 2 12

HM SGW 3 27

HM SOW 1 3

HM STW 1 3

HM/SW SGW 5 18

HM/SW SOW 1 2

SW GW(GROG) 2 34

SW OW(SHELL) 1 16

SW SGW 34 187

SW SOW 5 11

SW/WM SGW 12 59

WM GW(CALC) 4 8

WM GW(FINE) 2 5

WM GW(GROG) 3 10

WM SGW 39 207

Table 16. Early Roman fabric summary.

Romano-British (RB).

Although the range of this era has a wide

definition, (from C1st AD to early C5th AD),

it is typified by the appearance of late Roman

(C3rd AD to C4th AD) forms and fabrics.

Hand made and slow wheel made products

are extremely rare in this era; by far the

majority of pottery is being produced on fast

wheels.

Besides the dominant locally made Sandy

Grey Ware (SGW) and Sandy Orange Ware

(SOW), regional products are evident,

including the readily recognisable fabrics from

Horningsea; (HORNCW) and (SGW(HORN)),

although only as single sherds and the

ubiquitous late era fabrics from the Nene

Valley industry; (NV WH), (NVCC) and

(NVVV).

Evidence for trade with more distant

markets is shown by the identification of

sherds from Oxfordshire products in Sandy

Grey Ware (SGW) and Red Slipped (OX RS),

oxidised wares from Hadham in Hertfordshire

(HAD OX) and New Forest Colour Coated

(NFCC). There were also four sherds of

Samian, with examples from Central (SAM

CG) and Southern (SAM SG) Gaul, dating to

the C2nd AD to C3rd AD.

TYPE FABRIC FAMILY

QUANTITY WEIGHT (g)

HM GW(GROG) 1 3

HM HORN CW 1 13

HM SGW 1 9

SW SGW 4 37

SW SGW(HORN) 1 7

SW SOW 2 14

WM HAD OX 7 25

WM NFCC 1 1

WM NV WH 1 34

WM NVCC 4 20

WM NVVV 1 2

WM OW 1 2

WM OX RS 2 8

WM REDW 1 2

WM SAM 2 4

WM SAM CG 1 1

WM SAM EG 1 3

WM SGW 1 6

WM SGW 95 613

WM SOW 23 114

WM SREDW 9 35

WM STW 1 6

Table 17. Romano-British fabric summary.

Medieval (MED).

By the Medieval era, defined broadly as

C10th – C16th, hand made and slow wheel

made products return to prominence. Almost

all are sand tempered fabrics, with grey wares

making up the most numerous finds.

There was only one clearly identifiable

early Saxon sherd in an orange (OW) fabric


22

with organic (ORG) tempering, dating to C5th

– C7th. However, as noted above, two sherds

occurring in Grey Ware fabrics could have

originated in the early part of this era, but due

to their fragmentary nature, could not be

readily distinguished from Iron Age fabrics.

TYPE FABRIC FAMILY

QUANTITY WEIGHT (g)

HM OW 1 6

HM SGW 7 97

HM SREDW 3 140

HM/SW SGW 2 35

HM/SW SOW 1 7

SW SGW 9 67

SW SREDW 3 7

SW/WM SREDW 1 4

WM SGW 30 222

WM SREDW 6 20

Table 18. Medieval fabric summary.

One possible example of Ipswich Ware

(IPS) was identified dated C10th – C14th.

Glazed wares, mainly Green Glazed, make up

one third of the sherds from this era. 13 of the

63 sherds are recorded as Grimston–Type,

Sandy Glazed Ware (SGW); 5 in hand made

(HM) and 8 fast wheel made (WM) forms.

Jars and bowls are the dominant forms, but

there are also examples of jugs, dishes and

pitchers.

Post Medieval (PMED).

The range of fabrics and numbers of sherds

recorded for the Post Medieval era, are much

reduced. This may have been due to the

discard of clearly modern wares, such as

Transfer Prints, Porcelains and Stone Ware, at

the washing and sorting stages of finds

processing.

The period represented ranges from C15th

– C18th and forms are more utilitarian

comprising; jugs, bowls and dishes.

Slow wheel made products, from the earlier

part of the era, are still represented by around

one quarter of the sherds. These are mainly

Green Glazed Red (GRE) and Sandy Red

wares (SREDW).

A small number of late era, fast wheel

made jar sherds, in Sandy Red Ware fabrics,

were slipped. These included one which was

slipped and glazed and one jar sherd which

was red slipped.

TYPE FABRIC FAMILY

QUANTITY WEIGHT (g)

SW SREDW 9 43

WM SGW 3 39

WM SREDW 14 84

Table 19. Post Medieval fabric summary.

Conclusions.

As with the CBM, much of the pottery

assemblage recovered during the fieldwalking

exercise at Reach was very degraded. It had a

low average sherd weight of 7.1g (Table 12)

with, predictably, the more friable Prehistoric

and Saxon sherds having the lowest average

weights, albeit on low counts. Some of the

individual eras have somewhat higher average

weights, but are represented by low sherd

counts.

The pottery assemblage demonstrates how

activity at, and around the site, was largely

continuous from the prehistoric, right up to the

modern era, on the rich, fen-edge farmland.


23

Archiving

A copy of this report will be lodged with

the Cambridgeshire Historic Environment

Team, for inclusion in the Historic

Environment Record. It will also be made

available to download from the CAFG

website. The full recording spreadsheets for

the CBM and pottery may also be available

from the CAFG website, or by application to

the group.

A representative selection of CBM forms

and fabrics and also of the pottery will be

retained. In the short term, it will be held in

boxes in the general storage area of Oxford

Archaeology East’s premises at Trafalgar

Way, Bar Hill, Cambridgeshire and may

ultimately be deposited in Cambridgeshire

County Council’s ‘Deepstore’ facility in

Cheshire.

Acknowledgements

We would like to thank Alice Lyons, who

carried out the assessment, cataloguing and

dating of the pottery, which did much to bring

out the story of human activity at the Reach

villa. Thanks also are due to Carole Fletcher

(OAE) and Rob Atkins (now MOLA), whose

comments and suggestions have aided the

interpretation of the assemblage.

We are especially appreciative of the

continued support and access to facilities at

Bar Hill, by Oxford Archaeology East. Finally,

thanks go to those CAFG colleagues, who read

and commented on early drafts of this report

and all the other past and present members of

CAFG, without whose efforts and enthusiasm

the Reach project may never have been

completed.


24

Bibliography

AOC 2017. Cambridge Road, Melbourne, Cambridgeshire: An Archaeological Evaluation Report.

AOC Archaeology Group Twickenham. p102.

Atkinson, T.D. 1894. On a Roman House at Swaffham Prior, Cambridgeshire. Proceedings of the

Cambridge Antiquarian Society, 31st October 1892 – 17

th May 1893. N° XXXV. Vol. 8.2

(1894). pp.209 - 234

Betts, I.M. (1996). New Thoughts on Bricks With Sunken Margins. Information 68, July 1996.

pp.6 - 10. The British Brick Society. http://britishbricksoc.co.uk/wp-content/uploads/

2013/07/BBS_68_1996_July.pdf Accessed: 07/06/2020

BHO https://www.british-history.ac.uk/vch/cambs/vol10/pp225-227#fnn14:07/06/2020

Brodribb, G. 1987. Roman Brick and Tile: An Analytical Survey and Corpus of Surviving

Examples. Gloucester: Alan Sutton.

Coates, C. B. 2014. The Consumption and Distribution of Romano-British Ceramic Roofing

Materials in Cambridgeshire. Unpublished BA dissertation, University of Leicester.

Coates, C. B. 2015. Whole Way Cottage, Harlton, Cambridgeshire. 1994 (HAR003):

Assessment of the Roman Ceramic Building Material. Cambridge Archaeology Field Group.

CAFG (Forthcoming a). Manor Farm, Great Eversden, Cambridgeshire (GTE002):


CAFG (Forthcoming b). Cantelupe Farm, Haslingfield, Cambridgeshire (HAS06):


CAFG (Forthcoming c). Childerley, Cambridgeshire (CHLxxx):


CUCAP 1977. Cambridge University Collection of Aerial Photography.

http://www.cambridgeairphotos.com/location/ccj5/ Accessed: 19/09/2016

Fox, C. 1925. Excavations in the Cambridgeshire Dykes IV. The Devil’s Dyke: Excavations in 1923

and 1924. Proceedings of the Cambridge Antiquarian Society, October 1923 – May 1924.

Vol. XXVI. pp 90 – 127.

Lyons, A. 2019. Reach, Cambridgeshire: Pottery and CBM assessment, Microsoft Excel spreadsheet.

Lyons Archaeology for Cambridge Archaeology Field Group.

McComish, J. M. 2015. A Guide to Ceramic Building Materials. York Archaeological Trust.

http://www.yorkarchaeology.co.uk/wp-content/uploads/2015/08/

A-guide-to-ceramic-building-material.pdf Accessed: 22/02/2016

NHLE. National Heritage List for England. http://www.historicengland.org.uk/listing/the-list/

Sherlock, D. (1999). Brickmaking Accounts for Wisbech, 133 – 1356. Proceedings of the

Cambridge Antiquarian Society, Vol. LXXXVII for 1998. pp.59-70

Smith P.T. 2001. On Small Yellow bricks….from Holland. Construction History, Vol. 17. pp.31-42

Ramos Sáinz, M.L. 2003. Tejas romanas procedentes de San Juan de Maliaño (Cantabria):

Estudio analítico y experimental, en C.A.E.A.P. 25 años de investigaciones sobre

patrimonio cultural de Cantabria. ed. Ayto Camargo. pp.173-191.

http://personales.unican.es/ramosml/tegulas1.html

RCHM 1972. 'Reach', in An Inventory of the Historical Monuments in the County of

Cambridgeshire, Volume 2, North-East Cambridgeshire. (1972). London. pp. 85-90

http://www.british-history.ac.uk/rchme/cambs/vol2/pp85-90. Accessed: 18/05/2015.

http://britishbricksoc.co.uk/wp-


25

Webster, C.J. 1988. Ernest Greenfield’s Excavations at Exning Roman Villa. Proceedings of the

Cambridge Antiquarian Society, Vol. LXXVI for 1987. pp.41-66

Wessex 2011. Castor, Peterborough, Cambridgeshire: Archaeological Evaluation and Assessment of

Results. Wessex Archaeology, Salisbury, Wiltshire. Report reference: 74155.01

Warry, P. 2006a. Tegulae: Manufacture, Typology and Use in Roman Britain. British

Archaeological Reports, British Series. N°417.

Warry, P. 2006b. A Dated Typology for Roman Roof Tiles (Tegulae). Journal of Roman

Archaeology. Vol. 19, part 1, pp. 246-265.

Warry, P. 2010. Dinnington Ceramic Building Material and Stone Tile. https://www.academia.edu/

38762204/Dinnington_Ceramic_Building_Material_and_Stone_Tile_Overview

Accessed: 05/06/2020.


26

Appendix 1. Extent of fieldwalking grid (white squares), showing Roman CBM find spots.

AG

AF

AE

AD

AC

AB

AA

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41


27

Appendix 2. Extent of fieldwalking grid (white squares), showing Post Roman CBM find spots. AG

AF

AE

AD

AC

AB

AA

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41


28

APPENDIX 3. Pottery forms by era.

FORM

FORM BY ERA

PRE IA IA/ESAX LIA LIA/ER ER RB SAX LSAX/EMED MED PMED TOTALS

N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g)

AMPH 4 258 4 258

BEAK 1 5 3 5 4 10

BOWL 1 9 2 28 1 21 1 3 1 4 1 1 3 6 1 7 11 79

CORDONNED BOWL 1 5 1 5

CORDONNED JAR 1 4 6 58 7 62

DISH 4 39 6 45 1 4 2 36 4 53 17 177

DISH(REEDED RIM) 1 8 1 8

DISH(SINGLE GROOVE UNDER RIM) 1 7 1 7

DISH/LID 1 11 1 11

DISH/PLATTER 1 35 1 35

EVERTED RIM JAR 1 8 1 8

FDISH 2 13 2 13

FLAG 2 4 12 44 14 48

FLAG/BOWL 1 3 1 3

FLAGON 1 3 1 10 2 13

FLANGED BOWL 1 35 1 35

FLANGED DISH 1 18 1 18

FRAG 2 2 2 2

HIGH-SHOULDERED JAR 1 11 1 11

JAR 1 23 9 56 26 208 40 254 1 13 29 121 14 62 120 737

JAR WITH EVERTED RIM) 1 7 1 7

JAR(RILLED) 1 31 1 31

JAR(WITH LARGE EVERTED RIM) 1 30 1 30

JAR/BEAK 1 1 3 15 7 21 11 37

JAR/BOWL 6 49 1 10 24 223 30 157 50 380 36 216 15 126 2 12 164 1173


29

APPENDIX 3. Pottery forms by era (cont.).

FORM

FORM BY ERA

PRE IA IA/ESAX LIA LIA/ER ER RB SAX LSAX/EMED MED PMED

TOTALS

N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g) N°

Wt

(g)

JAR/SJAR 2 26 10 109 10 123 1 9 1 35 2 11 26 313

JUG 4 211 4 211

LID 1 7 1 3 2 10

MORT 1 8 1 8

MORT (REEDED RIM) 1 34 1 34

NECKLESS JAR WITH A FLAT RIM 1 10 1 10

PITCHER 1 24 1 24

PLATE/CUP (WALTERS 79 OR 80) 1 3 1 3

POT/CBM 1 3 1 9 2 12

POT/DAUB 2 7 2 7

PURN 1 31 1 31

SJAR 5 85 35 702 1 5 3 27 2 18 2 25 48 862

SJAR; LARGE EVERTED RIM 1 40 1 40

SJAR; WITH LARGE EVERTED RIM 1 53 1 53

WJAR 1 22 1 22

TOTALS 1 9 11 126 2 31 41 431 94 1315 100 800 131 967 1 4 1 13 57 596 25 166 464 4458


30

APPENDIX 4. Pottery finds distributions by era, per 10m grid square.

IRON AGE

LATE IRON AGE

A G

A F

A E

A D

A C

A B

A A

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1

A G

A F

A E

A D

A C

A B

A A

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1


31

APPENDIX 4. Pottery finds distributions by era, per 10m grid square (cont.).

LATE IRON AGE/EARLY ROMAN

EARLY ROMAN

A G

A F

A E

A D

A C

A B

A A

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1

A G

A F

A E

A D

A C

A B

A A

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1


32


ROMANO – BRITISH

MEDIEVAL

A G

A F

A E

A D

A C

A B

A A

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1

A G

A F

A E

A D

A C

A B

A A

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1


33


POST MEDIEVAL

A G

A F

A E

A D

A C

A B

A A

Z

Y

X

W

V

U

T

S

R

Q

P

O

N

M

L

K

J

I

H

G

F

E

D

C

B

A

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 2 0 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1 3 2 3 3 3 4 3 5 3 6 3 7 3 8 3 9 4 0 4 1


34

APPENDIX 5. Gridsquares by finds bags table.

BAG LETTER No

BAG LETTER No

BAG LETTER No

BAG LETTER No

BAG LETTER No

1 AF 3

49 O 1

100 P 9

146 R 17

195 A 21

2 AF 5

50 N 6

101 L 7

147 R 14

201 A 22

3 AD 6

51 N 3

102 K 8

148 S 18

202 A 27

4 AE 8

52 O 0

103 H 8

149 Q 18

204 A 24

5 AE 8

53 M 6

104 I 6

150 U 12

205 B 21

6 AD 6

55 N 5

105 F 7

151 P 10

206 C 19

7 AD 5

56 L 2

106 F 5

152 P 15

207 D 21

8 AE 3

57 L 1

107 C 7

153 P 13

208 E 19

9 AE 9

58 K 1

108 B 8

154 O 12

209 E 21

10 AD 5

59 J 5

109 D 9

155 O 10

210 D 24

11 AD 3

60 H 7

110 AE 10

156 O 9

211 E 25

12 AD 7

61 H 5

111 AD 11

157 O 18

212 C 25

13 AC 6

62 H 4

112 AC 10

158 O 17

213 C 26

14 AB 5

63 H 3

113 AA 10

159 N 16

214 C 28

15 AB 3

64 G 2

114 X 9

160 N 13

215 E 28

16 AA 4

65 I 0

115 AC 11

161 M 10

216 F 26

17 AB 8

66 G 0

116 AB 12

162 M 17

217 F 25

18 AA 8

67 J 3

117 AB 11

163 L 17

218 E 23

19 Z 8

68 E 4

118 AA 11

164 L 16

219 H 29

20 AA 7

69 E 2

119 AA 12

166 J 17

220 H 27

21 Z 5

71 E 1

120 AB 14

167 L 13

221 H 25

22 Z 3

72 E 0

121 AA 14

168 K 11

222 G 24

23 Y 6

74 C 0

122 AA 14

169 J 12

223 G 23

24 Y 7

75 B 2

123 AA 15

170 H 10

224 G 21

25 V 8

76 A 6

124 Z 16

171 G 11

225 H 20

27 Y 8

77 B 3

125 Y 17

172 I 14

226 H 19

28 X 8

78 C 4

126 Z 12

173 H 14

227 I 17

29 W 8

79 B 5

127 Y 12

174 I 12

228 J 19

30 W 6

80 AE 9

128 W 12

175 H 12

229 J 18

31 V 4

81 AE 7

129 X 10

176 H 15

230 J 21

32 W 1

82 AD 8

130 W 10

178 D 10

231 K 22

33 V 6

83 AC 9

131 W 13

179 D 11

232 K 18

34 T 6

84 AA 9

132 X 14

180 A 11

233 L 18

36 U 8

85 Z 7

133 X 15

182 E 15

234 L 19

37 U 1

86 Z 8

134 U 11

183 D 16

235 L 21

38 U 3

87 Y 9

135 V 12

184 D 13

236 M 19

39 T 5

88 X 9

136 V 13

185 D 15

237 I 23

40 T 4

90 W 9

137 V 17

186 C 17

238 I 25

41 S 2

91 V 6

138 W 15

187 C 16

239 I 27

42 S 0

93 V 9

139 U 15

188 F 10

241 K 23

43 T 6

94 U 9

140 T 11

189 B 14

242 J 25

44 S 8

95 U 7

141 S 11

190 A 17

243 K 27

45 S 6

96 S 9

142 T 12

191 B 13

245 K 26

46 R 8

97 R 7

143 T 13

192 A 14

246 K 25

47 Q 6

98 R 9

144 R 13

193 E 17

247 L 24

48 P 3

99 P 7

145 R 15

194 A 18

248 N 20


35

APPENDIX 4. Finds bags and gridsquares table (cont.).

BAG LETTER No

BAG LETTER No

BAG LETTER No

BAG LETTER No

249 N 18

295 V 21

349 F 40

398 U 32

250 O 18

296 V 30

352 F 36

399 S 30

251 P 19

297 U 30

353 F 37

400 S 29

252 Q 19

298 W 29

354 G 40

401 T 30

253 R 19

299 V 28

355 H 37

402 U 29

254 S 19

302 W 27

356 H 36

403 T 31

255 U 18

303 W 25

357 J 37

404 T 30

256 V 18

304 X 25

358 F 39

405 U 31

257 P 20

305 W 23

359 J 41

406 V 31

258 Q 20

306 X 22

360 J 40

407 U 30

259 R 20

307 X 21

361 J 35

408 V 30

260 S 20

308 X 20

362 K 37 261 N 21

309 Y 19

363 M 34

262 M 23

310 Z 17

364 M 35 263 N 23

316 A 29

365 M 36

264 O 25

317 A 30

366 K 40 265 M 27

318 A 35

367 L 41

266 N 27

319 B 32

368 M 41 267 N 29

320 B 34

369 N 34

268 Q 28

321 C 29

370 O 33 269 Q 27

322 C 30

371 O 31

270 Q 25

323 F 31

372 P 30 271 P 24

324 C 32

373 N 38

272 P 23

325 C 33

374 O 40 273 P 22

326 E 32

376 P 36

274 O 26

327 F 34

377 P 41 275 R 29

328 F 33

378 O 38

276 R 28

329 G 30

379 R 36 277 S 28

330 F 29

380 R 37

278 S 26

331 I 32

381 R 38 279 S 25

332 H 34

382 S 39

280 R 24

333 J 29

383 S 41 281 S 23

334 K 29

384 R 41

282 S 23

335 L 28

385 R 41 283 S 22

336 O 29

386 S 37

284 T 21

337 O 28

387 U 37 285 T 23

338 K 31

388 S 36

286 T 25

339 K 32

389 U 36 287 T 26

340 M 33

390 Q 33

288 U 29

341 M 32

391 R 31 289 T 29

343 C 35

392 R 33

290 U 28

344 C 37

393 S 35 291 V 27

345 D 39

394 T 33

292 V 26

346 D 40

395 S 33 293 V 24

347 B 38

396 S 31

294 V 22

348 A 36

397 T 31

Documents

Cambridge Archaeology Field Group July 2020